CLAVATA modulates auxin homeostasis and transport to regulate stem cell identity and plant shape in a moss

Summary The CLAVATA pathway is a key regulator of stem cell function in the multicellular shoot tips of Arabidopsis, where it acts via the WUSCHEL transcription factor to modulate hormone homeostasis. Broad‐scale evolutionary comparisons have shown that CLAVATA is a conserved regulator of land plant stem cell function, but CLAVATA acts independently of WUSCHEL‐like (WOX) proteins in bryophytes. The relationship between CLAVATA, hormone homeostasis and the evolution of land plant stem cell functions is unknown. Here we show that in the moss, Physcomitrella (Physcomitrium patens), CLAVATA affects stem cell activity by modulating hormone homeostasis. CLAVATA pathway genes are expressed in the tip cells of filamentous tissues, regulating cell identity, filament branching, plant spread and auxin synthesis. The receptor‐like kinase PpRPK2 plays the major role, and Pprpk2 mutants have abnormal responses to cytokinin, auxin and auxin transport inhibition, and show reduced expression of PIN auxin transporters. We propose a model whereby PpRPK2 modulates auxin gradients in filaments to determine stem cell identity and overall plant form. Our data indicate that CLAVATA‐mediated auxin homeostasis is a fundamental property of plant stem cell function, probably exhibited by the last shared common ancestor of land plants.


Introduction
Organ size and shoot architecture are determined by the number and activity of stem cells in the growing shoot tips of flowering plants such as Arabidopsis (Fletcher, 2018). The size and integrity of Arabidopsis shoot tips is maintained by the action of a molecular feedback loop involving CLAVATA peptides and receptors and the WUSCHEL transcription factor (Schoof et al., 2000;Somssich et al., 2016). CLAVATA3 is expressed in stem cells in the outermost cell layers (Fletcher et al., 1999), encoding a protein that is processed to form a small diffusible peptide (Rojo et al., 2002;Lenhard & Laux, 2003;Kondo et al., 2006;Ohyama et al., 2009). The CLAVATA3 peptide acts as a ligand to the CLAVATA1 receptor which is active in inner cell layers of the shoot tip (Clark et al., 1997;Ogawa et al., 2008), and signalling via CLAVATA1 confines the expression of WUSCHEL to a few cells at the centre of the shoot tip (Schoof et al., 2000). In turn, the WUSCHEL protein moves to the outermost cell layers of the shoot tip (Yadav et al., 2011;Daum et al., 2014), promoting expression of CLAVATA3. Mutants with defective CLAVATA or WUSCHEL function respectively over-proliferate cells in the shoot tips or are unable to maintain the stem cell population, resulting in enlargement of the tips or shoot termination (Laux et al., 1996;Clark et al., 1997). CLAVATA function depends on the maintenance of low levels of auxin signalling permissive to stem cell activity in the central zone of the shoot tips by WUSCHEL (Schoof et al., 2000;Ma et al., 2019).
Unlike Arabidopsis, the growing tips of mosses such as Physcomitrium patens comprise a single apical cell (Harrison, 2017). Spores germinate to form a branching mat of filamentous tissue termed the protonema, and tip growth in the apical cells of each protonemal filament extend plant spread (Menand et al., 2007a). When growth commences, filaments have chloronemal identity and serve a primarily photosynthetic function, having many large dark green chloroplasts, but later on, more rapidly growing foraging filaments with smaller pale chloroplasts develop (caulonemata) (Ashton et al., 1979). Whereas chloronemal apical cells cleave in a plane perpendicular to the main axis of growth, caulonemal apical cells cleave obliquely and generate cells that are longer than chloronemal cells (Ashton et al., 1979). The relative growth of chloronemata and caulonemata determines the size and shape of the plant, such that plants comprising only chloronemata are small and round, whereas plants with predominantly caulonemata are larger and have an irregular foraging fringe (Prigge et al., 2010).
There are three P. patens WUSCHEL-like homeobox (WOX) genes, and previous work detected constitutive expression of PpWOX13LA and PpWOX13LB, but no expression of PpWOX13LC; PpWOX13LA and PpWOX13LB protein fusions showed elevated expression in protonemal stem cells and stem cells forming in a leaf regeneration assay (Sakakibara et al., 2014). Ppwox13lab mutants were unable to initiate growth in leaf regeneration assays but were otherwise indistinguishable from wild-type plants during gametophyte development, and this divergence in phenotypes suggests that P. patens WOX and CLAVATA genes act independently (Sakakibara et al., 2014). Marchantia polymorpha CLAVATA pathway components (MpCLE2 and MpCLV1) likewise act in a WOX-independent manner (Hirakawa et al., 2020). Thus, whilst CLAVATA is a conserved regulator of land plant stem cell function, WOX function appears inessential in bryophytes, raising questions about the evolution of the CLAVATA-WUSCHEL gene regulatory network and the regulation of stem cell function in bryophytes and the last shared common ancestor of land plants. Here we show that CLAVATA acts in protonemata to repress the chloronema to caulonema developmental transition and propose a model whereby CLAVATA regulates stem cell identity by modulating auxin homeostasis and PIN-mediated auxin transport.

Research
New Phytologist were prepared in 70% ethanol. 100 mM stocks of L-Kynurenine (L-Kyn) were prepared using dimethyl sulfoxide (DMSO) as a solvent. 5 mM N-1-naphthylphthalamic acid (NPA) stocks were prepared in 1 ml DMSO made up to 50 ml with 70% EtOH. All reagents were added to warm growth media before pouring plates.

Phenotype analysis
Plant areas and perimeters were measured using FIJI from images taken using a VHX-1000 microscope (Keyence, Osaka, Japan) with a 910 objective, excluding gametophores. These values were used to calculate the perimeter ratio, the ratio between the measured perimeter and the perimeter of a perfectly circular plant of the same area. For cell identity measurements, filaments protruding from the margins of 4-wk-old plants were dissected and stained with 0.3% Toluidine Blue for 2 min, rinsed in water and mounted on slides with coverslips before imaging with a DM2000 microscope (Leica, Wetzlar, Germany) using a 940 objective or a VHX-1000 microscope (Keyence) using a 950-200 objective. The length and cell division angle of subapical cells of main filaments and of the second cell in branches with at least three cells were measured using FIJI as described previously (Coudert et al., 2019).

Generation of promoter::NGG lines
Promoter sequences from PpCLE3 (2427 bp), PpCLE4 (2867 bp), PpCLE5 (1731 bp) and PpCLE6 (1458 bp) were PCR-amplified using a proofreading polymerase and cloned directly into the SmaI site of a modified PIG1NGGII (Ishikawa et al., 2011) vector in which an NptII resistance cassette was substituted for the BSD cassette (Whitewoods et al., 2018). The promoter plus the first few amino acids of the peptide coding sequence were PCR-amplified from PpCLE8 (3216 bp) and PpCLE9 (1799 bp) before insertion into PIG1NGGII to generate a translational fusion with the reporter gene. All constructs were linearized with PmeI before plant transformation. Lines were screened using a forward primer from the PIG1 targeting locus and a promoter-specific reverse screening primer to check the 5 0 integration site, and a CaMV terminator forward primer and reverse primer from the PIG1 locus were used to check the 3 0 integration site. Southern analyses to verify targeting were undertaken using either a green fluorescent protein (GFP)-bglucuronidase (GUS) probe PCR-amplified to incorporate a DIG label (PpCLE3, PpCLE4, PpCLE5, PpCLE6 and PpCLE9) or a probe against the 35S::NptII resistance cassette (PpCLE8) as illustrated in Supporting Information Fig. S1, and using methods described in Whitewoods et al., (2018). Primer sequences are listed in Table S1.

Nucleic acid extraction
DNA for PCR and Southern analysis was extracted using a modified CTAB protocol (Doyle & Doyle, 1990). RNA for expression analyses was extracted from 5 d-old protonemata using a RNeasy Plant Mini Kit (Qiagen). Genomic DNA removal and cDNA synthesis were performed with a Quantitect Reverse Transcription kit (Qiagen).

Expression analysis
Reverse transcription (RT)-PCR was performed using a 1 : 10 dilution of cDNA as a template, and EcoTaq polymerase (Desai & Pfaffle, 1995). Where possible, primers were designed to span an intron or were placed on an intron-exon boundary, and a UBIQUITIN transcript was amplified as a positive control (Table S2). Quantitative (q)PCR was performed using a SYBR Green Quantitect kit and a Stratagene M93005P bioanalyser with 95°C for 15 min, and then 94°C for 15 s, 60°C for 30 s, 72°C for 30 s cycling conditions for 40 cycles. Amplicon size was checked by dissociation curve. The efficiency of each primer pair was calculated on serial dilutions of WT cDNA, and only primer pairs with an efficiency between 90% and 110% were used in further experiments. To calculate fold-change for each sample relative to the WT expression levels, the DDCt method was used (Livak & Schmittgen, 2001). PINA, PINB, PINC, PIND and 60S genes were amplified using primers listed in Table S3.
Caulonemal filaments extending from the foraging fringe were dissected to observe expression at later developmental stages (Fig. 1d). No signal was detected in PpCLE1::NGG, PpCLE8::NGG or PpCLV1a::NGG lines, but the promoters of all remaining PpCLEs were active in caulonemal tip cells, and PpCLE5::NGG and PpCLE9::NGG lines showed the strongest signal. PpCLV1b::NGG lines frequently showed signal in side branch initials and the second and third cells from the tip, and PpRPK2::NGG lines showed strong signal in the majority of caulonemal tip cells. Taken together these expression data indicate roles for PpCLE5, PpCLE9 and PpRPK2 in CLE production and perception in caulonemal tip cells, and to a lesser extent indicate potential roles for PpCLV1b in caulonemal tip cells and branch initials.

P. patens clavata mutants have cell identity and plant shape defects
In order to investigate roles for CLAVATA in protonemal development, we first quantified overall plant spread in WT and mutant plants (Fig. 2a,b). We also quantified a measure of plant shape, the perimeter ratio, which reflects the circularity of plant spread. A perimeter ratio of 1 corresponds to a perfectly circular shape, and higher values indicate irregular plant shapes with no increase in area and increased production of caulonemata ( Fig. 2b). Whilst PpcleAmiR1-3 mutants had a similar size and shape to WT plants, PpcleAmiR4-7 mutants had greater spread but similar perimeter ratios to the WT, implying increased but uniform protonemal growth (Fig. 2b). Amongst receptor mutants, Ppclv1a, Ppclv1b and Ppclv1a1b mutants sometimes had increased areas and perimeter ratios, but the size and shape differences from WT plants were subtle and variable between experimental replicates. By contrast, Pprpk2 and Ppclv1a1brpk2 mutants were consistently and significantly larger with higher perimeter ratios than WT plants, indicating higher caulonema production and a key role for PpRPK2 in plant size and shape determination (Fig. 2b).
In order to further investigate the cellular basis of plant spread and perimeter ratio differences between lines, protonemal morphology was quantified by measuring the length and cell division plane orientations of subapical cells of main filaments (subapical cells) and side branches (branch cells), as these report differences between caulonemal and chloronemal identity (Coudert et al., 2019) (Fig. 2c,d). Ppclv1a, Ppclv1b and Ppclv1a1b mutants showed no significant differences from WT plants with respect to subapical and branch cell length or cell division plane orientation in most experimental replicates (Fig. S3). However, although PpcleAmiR1-3 mutants were the same overall size as wild-type plants, their cell types showed mixed chloronemal and caulonemal identities, with less oblique cell divisions in subapical cells and longer branch cells than WT plants (Fig. S3). PpcleAmiR4-7 mutant protonemata retained distinct cell identities but had longer subapical and branch cells than WT plants. Pprpk2 and sometimes

New Phytologist
Ppclv1a1brpk2 mutants had more distinct cell identities than WT plants, with longer and more oblique cell division plane orientations (Fig. S3).
More subtle differences in protonemal morphology also were observed following filament dissection (Fig. 2c). Wild-type filaments initiate branches from the second subapical cell, and branch growth continues to give protonemata regular branching patterns (Fig. 2c). Branches close to the tip were longer in PpcleAmiR1-3 mutants than WT branches in equivalent positions, and Pprpk2 and Ppclv1a1brpk2 mutants conversely had shorter branches (Fig. 2c, S4). Thus, CLE signalling regulates protonemal morphology in P. patens and PpRPK2 holds the main role as a repressor of plant spread, perimeter ratio and caulonemal identity.
Pprpk2 mutants are hypersensitive to cytokinin application As Ppclv1a and Ppclv1b mutant phenotypes were variable, and PpcleAmiR mutant phenotypes reflect changes in expression of several PpCLEs (Fig. S5), we focused further functional analyses on Ppclv1a1b, Pprpk2 and Ppclv1a1brpk2 mutants. In WT plants, filament identity reflects a hormonal interplay between auxin and cytokinin, and cytokinin suppresses caulonemal identity (Ashton et al., 1979). We therefore reasoned that PpRPK2 could promote cytokinin biosynthesis or enhance plants' response to cytokinin. To test the hypothesis that PpRPK2 promotes cytokinin biosynthesis, we used LC-MS/MS to quantify cytokinin concentrations in protonemal tissue from Ppclv1a1b, Pprpk2 and Ppclv1a1brpk2 mutants and WT controls. In flowering plants, cytokinin bases, and to a lesser extent ribosides are the active forms, whereas nucleotides and O-glucosides (OG) are inactive and can function in storage (Kieber & Schaller, 2018). In P. patens isopentenyl adenine (iP), trans-zeatin (tZ) and their corresponding ribosides were shown to have the highest biological activity in a bud induction assay (von Schwartzenberg et al., 2007), whereas neither ciszeatin (cZ), cis-zeatin riboside (cZR) nor any of the ribotides had a bud inductive role. Twenty-six types of cytokinin were assayed, including free bases as iP, tZ, cZ and dihydrozeatin (DHZ), and their ribotide (MP), riboside (R) and glycoside derivatives (Fig. S6). Thirteen types of cytokinins were detected, and although some mutant-specific differences were present at the level of conjugates such as cZRMP, tZOG and tZROG, the only free base present at higher concentrations in mutants was cZ in Ppclv1a1b mutants. No difference in the overall level of cytokinin was detected. Thus, the hypothesis that PpRPK2 reduces plant spread by promoting cytokinin biosynthesis was rejected.
We next tested the hypothesis that PpRPK2 enhances plants' response to cytokinin. To this end, we grew mutants and WT plants on media containing a solvent control (0.07% EtOH) or 100 nM of the synthetic cytokinin 6-benzylaminopurine (BAP), and analyzed their phenotypes after 4-5 wk of growth (Fig. 3). As expected, WT plants showed a decrease in plant spread in response to exogenously applied BAP (Fig. 3a,b). On average, the area of WT plants grown on 100 nM BAP was 26% AE 11% smaller than that of control plants, with no difference in perimeter ratio. The response to BAP treatment was sometimes weaker in PpcleAmiR1-3 and Ppclv1a1b mutants than in controls, but Pprpk2 and Ppclv1a1brpk2 plants grown on 100 nM BAP consistently showed an enhanced response, with a 59% AE 3% and 43% AE 9% (respectively) decrease in plant spread and reductions in perimeter ratio (Fig. 3a,b).
In order to study cytokinin responses at the cellular level, we dissected foraging filaments from plants and measured cell lengths and division planes as described previously (Fig. 3c). In WT plants, BAP treatment reduced cell lengths in both subapical and branch cells, whereas division plane angles were unaffected. In Ppclv1a1b and PpcleAmiR1-3 mutant subapical cells, responses to cytokinin were respectively similar to or weaker than WT responses. However, PpcleAmiR4-7, Pprpk2 and Ppclv1a1brpk2 subapical cells showed a large reduction in both cell length and division plane angle following BAP treatment. Thus, Pprpk2 and Ppclv1a1brpk2 mutant plants have an enhanced response to cytokinin in caulonemal tip cells as well as in whole plants, refuting the hypothesis that PpRPK2 promotes plants' response to cytokinin. Pprpk2 mutants showed further evidence of cytokinin hypersensitivity in gametophore initiation, which is normally upregulated by cytokinin (Ashton et al., 1979). Whilst WT plants grown on BAP showed a higher frequency of gametophore initiation than untreated controls, Pprpk2 mutants grown on BAP showed almost constitutive gametophore initiation (Fig. 3d).

Auxin synthesis is not elevated in Pprpk2 mutants, but mutant increases in plant spread require auxin
Because auxin can enhance or suppress cytokinin activity in P. patens (Ashton et al., 1979), we hypothesized that cytokinin hyper-sensitivity in Pprpk2 and Ppclv1a1brpk2 mutants could reflect differences in auxin biosynthesis. To test the dependence of clavata mutant phenotypes on auxin biosynthesis, we grew WT and mutant plants on media containing a pharmacological inhibitor of auxin synthesis (10 µM L-Kyn) or a solvent control (0.01% DMSO) (Fig. 4a-c). When grown on L-Kyn, WT plants showed significantly smaller areas than plants grown on a solvent control. Area decreased in all clavata mutants except Ppclv1b, and decreases were largest in PpcleAmiR4-7, Pprpk2 and Ppclv1a1brpk2 mutants (Fig. 4b). Perimeter ratios were consistently unaffected by L-Kyn in WT, PpcleAmiR1-3, PpcleAmiR4-7 and Ppclv1a1b plants. Whilst Pprpk2 and Ppclv1a1brpk2 mutant perimeter ratios remained higher than WT values, perimeter ratios decreased significantly following L-Kyn treatment in these mutants, indicating that their irregular shapes depend on auxin synthesis. To test the effect of L-Kyn on cell identity, we dissected foraging filaments and measured cell length and cell division plane orientations. While no response to L-Kyn was detected in branch cells, subapical cells of all lines were shorter and/or had division plane angles closer to 90°when plants were grown on L-Kyn, and these differences were significant in PpcleAmiR4-7, Ppclv1a1b and Pprpk2 mutants. Hence, WT and clavata mutant plants have a qualitatively similar response to auxin biosynthesis inhibition, and auxin synthesis is needed for the caulonemal overproliferation phenotype of Pprpk2 and Ppclv1a1brpk2 mutants.   As auxin promotes caulonemal development, we hypothesized that higher levels of auxin production could account for Pprpk2 and Ppclv1a1brpk2 mutant phenotypes. We therefore quantified biologically active auxin IAA, and its precursors anthranilate (ANT); L-tryptophan (TRP); and indole-3-pyruvic acid (IPyA) and degradation products IAA-glutamate (IAGlu) and 2oxindole-3-acetic acid (OxIAA) in protonemal tissues of WT and mutant plants. However, we found that Ppclv1a1b and four of five replicates of Pprpk2 tissue batches in fact had lower IAA concentrations than WT protonemata, whereas Ppclv1a1brpk2 IAA concentrations were variable (Fig. 4d). Ppclv1a1b and Pprpk2 protonemata contained more ANT than WT protonemata, and Ppclv1a1b and Ppclv1a1brpk2 mutant protonemata contained less OxIAA than WT samples. As Ppclv1a1b (and usually Pprpk2 mutant) protonemata contained less biologically active auxin (IAA) than WT protonemata, we rejected our hypothesis that higher levels of auxin synthesis contribute to Pprpk2 and Ppclv1a1brpk2 mutant phenotypes.

Pprpk2 mutants show abnormal developmental responses to auxin
Enhanced sensitivity to auxin could yield similar developmental outcomes to high auxin concentrations. To evaluate the response of clavata mutants to exogenous auxin, we grew plants on media containing either the synthetic auxin 1-naphthaleneacetic acid (1 µM NAA) or a solvent control (0.07% EtOH) (Fig. 5a-c). In line with previous studies (Ashton et al., 1979;Lavy et al., 2016), WT plant areas increased when grown with additional auxin, but perimeter ratios were unaffected. Although PpcleAmiR1-3, PpcleAmiR4-7 and Ppclv1a1b plants showed the same response as WT plants, Pprpk2 plants conversely showed both area and perimeter ratio decreases (Fig. 5a,b). At the cellular level, WT subapical cells showed a strong response to NAA application with division angles close to 90°. PpcleAmiR4-7 and Ppclv1a1brpk2 mutants also showed significant reductions in division angle following auxin application, and Ppclv1a1b and Pprpk2 mutants showed significant reduction in subapical cell lengths.
In order to further dissect the effects of auxin synthesis and auxin response, we uncoupled these two processes by growing plants on media containing either a solvent control (0.01% DMSO + 0.07% EtOH), 10 µM L-Kyn (reduced synthesis), 1 µM NAA (response plus endogenous synthesis), or a combination of 10 µM L-Kyn and 1 µM NAA (response in absence of synthesis). We reasoned that if mutant phenotypes are caused by enhanced auxin perception or response, mutants would respond in a similar way to saturating concentrations of exogenously applied auxin (1 µM NAA treatment) both with and without endogenous synthesis (combined treatment). Wild-type plant spread increased in L-Kyn + NAA treated plants relative to L-Kyn treated plants, illustrating the normal response to exogenously applied auxin in the absence of auxin synthesis (Fig. 5d,e). Whilst responses to exogenous auxin in Ppclv1a1b and Ppclv1a1brpk2 mutants were variable between experimental replicates, Pprpk2 mutants grown on the combined treatment had reduced areas relative to either single treatment (Fig. 5d,e). These data did not fit our hypothesis that Pprpk2 mutants should respond similarly to exogenously applied auxin regardless of the concentration of endogenous auxin synthesis. Thus, increased auxin responsiveness was insufficient to fully account for the phenotype of Pprpk2 mutants.
Pprpk2 mutants are sensitive to auxin transport inhibition and have low PIN expression levels The combined L-Kyn + NAA treatment above would not only change the amount of auxin present, but also its spatial distribution. Therefore, we next hypothesized that local auxin gradients could be important in Pprpk2 mutant phenotype determination. Promoter fusions have shown that PIN auxin transporter genes are highly expressed in protonemal tip cells of P. patens, and the chloronemal-to-caulonemal transition is accelerated in pina pinb mutants but suppressed in PINA and PINB (and to a lesser extent PINC) overexpressors (Viaene et al., 2014). To investigate a potential contribution of auxin transport to Pprpk2 mutant phenotypes, we first took a pharmacological approach, assaying the sensitivity of mutants to transport inhibition. We used naphthylphthalamic acid (NPA) as NPA strongly inhibits PIN function in Arabidopsis and PIN function is conserved (Geldner et al., 2001;Abas et al., 2021). We grew WT plants, Ppclv1a1b, Pprpk2 and Ppclv1a1brpk2 mutants on media containing a solvent control, 5 µM NPA, 1 µM NAA, or a combination of 5 µM (a) Micrographs of 4-wk-old plants grown on media containing a solvent control (EtOH) or auxin (1 µM 1-naphthaleneacetic acid (NAA)). Bar, 1 mm. (b) Quantitative analyses showed that Pprpk2 and Ppclv1a1brpk2 mutant plants show an auxin-dependent decrease in plant spread and perimeter ratio, whilst all other backgrounds increased in area in response to auxin treatment and showed no change in perimeter ratio (two-way ANOVA and Tukey's honestly significant difference (HSD) test. n ≥ 30; *, P < 0.05 between treatment and control). (c) Wild-type (WT), PpcleAmiR4-7 and Ppclv1a1brpk2 subapical cell division angles diminished following 1 µM NAA treatment, and Ppclv1a1b and Pprpk2 subapical cell lengths decreased, but branch cells and PpcleAmiR1-3 mutant cells showed no change in length or cell division plane angle. Letters indicate significant differences between groups, error bars indicate SD (multi-way ANOVA and Tukey's HSD test; n ≥ 90 for all other genotypes; P < 0.05). (d) Micrographs of 4-wk-old plants grown on media containing a solvent control (0.07% EtOH + 0.01% DMSO), 10 µM L-Kynurenine (L-Kyn), 1 µM NAA or a combination of 10 µM L-Kyn and 1 µM NAA. Bar, 5 mm. (e) Quantitative analyses showed that whilst WT plant spread showed little response to exogenously applied auxin (1 µM NAA), auxin synthesis inhibitors (10 µM L-Kyn) or a combination of 1 µM NAA and 10 µM L-Kyn, Pprpk2 mutant spread strongly decreased in all treatments and the combined treatment led to the strongest decrease. The response of Ppclv1a1b and Ppclv1a1brpk2 mutants varied between experimental replicates, as did perimeter ratios. Asterisks above data indicate significant difference from untreated controls, asterisks above bars indicate significant difference between treatments of interest (n ≥ 30; one-way ANOVA and Tukey's HSD test on each genotype; P < 0.05). In boxplots, horizontal lines represent median values, boxes represent the interquartile range, whiskers represent largest and smallest values within 1.59 above or below 75 th and 25 th percentiles, respectively, and black circles represent outliers.

Research
New Phytologist NPA and 1 µM NAA (Fig. 6a,b). As expected, WT plant spread increased following treatment with auxin in the single and combined treatments, but perimeter ratios decreased following NPA treatment. Whilst Ppclv1a1b mutants showed a similar response to WT plants, Pprpk2 and Ppclv1a1brpk2 mutants conversely showed mild growth suppression following treatment with 5 µM NPA or 1 µM NAA, and strong growth suppression following the combined treatment. We therefore concluded that Pprpk2 and Ppclv1a1brpk2 mutants were hypersensitive to auxin transport inhibition. To assay the molecular basis of sensitivity to auxin transport inhibition, we quantified PIN expression in protonemal tissues by qPCR (Fig. 6c). Whilst expression levels of the noncanonical PIND did not differ between WT and mutant samples, PINA expression was lower in Pprpk2 mutants and PINB expression was lower in all mutants than in WT tissue (Fig. 6c).
PINC is expressed at around the PCR detection limit in WT protonemata (Bennett et al., 2014), and we were unable to detect expression in any mutant.

Discussion
The data above led us to a model whereby CLAVATA controls auxin biosynthesis and the expression of PINA-C, but Physcomitrium patens RECEPTOR-LIKE KINASE (PpRPK2) has the major role in suppressing the chloronema-to-caulonema developmental transition (Fig. 6d). Pprpk2 and Ppclv1a1b mutants show similar reductions in auxin concentrations and PIN expression (Figs 4d, 6c), but Pprpk2 mutants show a much more severe mutant phenotype relative to wild-type (WT) plants (Fig. 2). As auxin synthesis promotes but auxin transport represses caulonemal development (Ashton et al., 1979;Viaene et al., 2014;Lavy et al., 2016), and CLAVATA promotes auxin synthesis and transport, our data point to a stronger role for auxin transport than synthesis in specifying caulonemal identity. PpRPK2 is most strongly expressed in tip cells, but PpCLV1a and PpCLV1b are respectively expressed at low levels or more strongly away from the tip (Fig. 1d). We propose that these differences in gene expression contribute to differences between Pprpk2 and Ppclv1a1brpk2 and Ppclv1a1b mutant phenotypes, and that PpRPK2 controls the auxin transport status of tip cells to affect plant growth (Fig. 6e). An altered auxin distribution in the tip cells of Pprpk2 mutants or subapical cells of Ppclv1a1b mutants would support our hypothesis. An alternative hypothesis is that PpCLV1a and PpCLV1b could act at least partially independently of PpRPK2, contributing to observed differences in mutant phenotypes between Ppcvl1a1b and Pprpk2 mutants, and there is some support for this hypothesis in recently published work (Cammarata et al., 2021;Demko et al., 2021).
Although protonemal tips cells are the distal site of phenotype determination, auxin signalling is concentrated at the centre of plants (proximally) (Menand et al., 2007b;Jang & Dolan, 2011). Gain-of-function pACT::PpPINGFP mutants do not produce caulonemata, and PINs are expressed most strongly in protonemal tip cells and a few subapical cells (Viaene et al., 2014). Our model for PpRPK2 function fits with these prior data by suggesting PIN can respond strongly to auxin if repressor ARF (PpARFb) levels are low, and PpARFb depletion by tasiRNAs in a subset of filaments patterns caulonema initiation and protonemal branching at plants' foraging fringe (Plavskin et al., 2016). PpCLE and PpRPK2 expression are highest in caulonemal tip cells, and as CLE peptides are diffusible, PpRPK2 could act in parallel or in sequence with PpARFb to pattern caulonema initiation (Fig. S7). As CLAVATA genes are not expressed in secondary chloronemata, we propose that differences in the auxin response between chloronemata and caulonemata in WT plants could reflect distinct domains of CLAVATA and PpARFb activity.
From an evolutionary perspective, our work suggests that roles for CLAVATA in modulating PIN activity, auxin response and stem cell identity (Pallakies & Simon, 2014;Han & Hwang, 2018;Racolta et al., 2018) are conserved within the land plants. In Arabidopsis, the WUSCHEL transcription factor acts downstream of CLAVATA to maintain low auxin conccentrations in stem cells of the central meristem zone. WUSCHEL belongs to the T3 clade of the WUSCHEL-LIKE HOMEOBOX (WOX) gene family, and T3 WOXes arose from a (T2 + T3) WOX gene duplication pre-dating the origin of vascular plants (Wu et al., 2019). Bryophytes lost the T2/T3 WOX precursor lineage (Wu et al., 2019), and remaining T1 WOX genes act independently of CLAVATA in bryophytes (Sakakibara et al., 2014;Hirakawa et al., 2020). Thus, it is likely that other transcription factors act in place of WUSCHEL to regulate PIN expression in P. patens. In Arabidopsis, the TDIF/PXY module regulates ARF action independently of WOX genes (Han & Hwang, 2018), and we propose that PpRPK2 could act via an ARF intermediary. Our data suggest that downstream components of the CLAVATA gene regulatory network have been remodelled during evolution. WUSCHEL could have been co-opted into the CLAVATA gene regulatory network in euphyllophytes, or a T3 WUSCHEL precursor could have been co-opted into the CLAVATA gene regulatory network in the last common ancestor of vascular plants. Alternatively, CLAVATA could have acted via a T2/T3 WOX precursor to regulate auxin homeostasis in the last common ancestor of land plants. Fig. 6 PIN-mediated auxin transport is dampened in Physcomitrium patens clavata mutants. (a) Images of 4-wk-old plants grown on a solvent control, 5 µM N-1-naphthylphthalamic acid (NPA), 1 µM 1-naphthaleneacetic acid (NAA) or a combination of 5 µM NPA and 1 µM NAA. Bar, 5 mm. (b) Quantitative analyses showed that Pprpk2 and Ppclv1a1brpk2 mutant plants showed auxin-and auxin transport-dependent decreases in plant spread and perimeter ratio, whereas wild-type (WT) and Ppclv1a1b plants showed auxin dependent increases in plant spread and transport-dependent reductions in perimeter ratio (CLV, CLAVATA; RPK, RECEPTOR-LIKE KINASE). In the boxplot, horizontal lines represent median values, boxes represent the interquartile range, whiskers represent largest and smallest values within 1.59 above or below 75 th and 25 th percentiles, respectively, and black circles represent outliers. Asterisks above data indicate significant difference from untreated plants of the same genotype, asterisks above bars indicate significant difference between treatments of interest (n ≥ 30; one-way ANOVA and Tukey's honestly significant difference (HSD) test on each genotype; P < 0.05). (c) Quantitative (q)PCR showed that PINA expression was depressed in Pprpk2 mutants relative to WT plants, and PINB and PINC expression were depressed in all mutants. PIND expression showed similar expression levels in WT plants and mutants. 60S was used as housekeeping control (n = 3; n.d., not determined; *, P < 0.05 between WT and mutant samples. No significant differences were found between mutants).

Author contributions
ZNV and JH conceived the project and designed the experiments; AC and YK engineered promoter::NGG fusion lines as part of the Leeds Moss Transformation Service; ON and AP performed hormonal quantification and data analysis. All remaining experimental work was performed by ZNV with help from CM, WL and AS and supervision from JH; ZNV and JH analysed the data, wrote the manuscript draft and incorporated feedback from all authors.

Supporting Information
Additional Supporting Information may be found online in the Supporting Information section at the end of the article.       Table S1 List of primers used for promoter::NGG line generation.

Table S2
List of primers used for reverse transcription polymerase chain reaction (RT-PCR) in PpcleAmiR lines (Fig. S5).